Convalescent sera from proven cases of infection with Mycoplasma pneumoniae, and rabbit antisera to M. pneumoniae and to human erythrocyte glycoprotein contained cold hemagglutinins which were reactive only for human erythrocytes. Only the human serum cold agglutinins were inhibited by soluble integral glycoproteins derived from human erythrocyte ghosts by treatment with chloroform-methanol. Rabbit antiserum to chloroform-methanol glycoprotein, as well as to M. pneumoniae, fixed complement with either M. pneumoniae or chloroform-methanol glycoprotein antigens. The findings support the hypothesis that the cold agglutinins elicited by M. pneumoniae infection represent a cross-reaction between determinants common to erythrocyte glycoprotein containing I antigen and the membrane of M. pneumoniae.Cold agglutinins (CA) which appear in the serum during the course of human infection with Mycoplasma pneumoniae have specificity for the I antigen, the determinants) of which involve N-acetylneuraminic acid (6,20). The I specificity is detectable in glycoproteins solubilized from erythrocyte membranes by any of several procedures (15) and in various body fluids. These soluble I-like substances display considerable heterogeneity as judged by immunochemical analysis with anti-I sera from patients with the chronic "cold agglutinin syndrome" unrelated to infection with M. pneumoniae. The mechanism by which CA are elicited by M. pneumoniae has not yet been clearly defined. We present further evidence which suggests that early (immunoglobulin M [IgM]) CA antibody evoked by M. pneumoniae is crossreactive with native erythrocyte antigens which share determinants with M. pneumoniae. MATERIALS AND METHODS M. pneumoniae, The FH strain of M. pneumoniae was obtained through the courtesy of William Mogabgab, Tulane University School of Medicine. It had been isolated in embryonated eggs from the sputum of a patient with CA-positive pneumonia (18) and was subsequently maintained in broth and agar media as described by Chanock et al. (2). The organism was grown in bulk by the method of Somerson et al. (23) in 5-liter Povitsky bottles containing 300 ml of broth.After inoculation, bottles were incubated at 37°C until a confluent sheet of colonies had attached to the surface, usually within a period of 3 to 5 days, during which the pH of the medium dropped to ca. 6.5. The broth was decanted and discarded. The sheet of organisms adherent to the glass was washed three times with phosphate-buffered saline, pH 7.2 and scraped into phosphate-buffered saline with a rubber policeman. The organisms were washed three times more in phosphate-buffered saline by alternate centrifugation at 800 x g for 30 min and resuspension, and finally resuspended to the desired concentration and stored at -70'C until used. Colony counts of M. pneumoniae were done by the method of Kim et al. (14).Antigens for complement fixation (CF) tests. Water-soluble glycoproteins were obtained by treatment of human erythrocyte ghosts with chloroform-methanol (CM) (9). The product of...

Respiratory infection with Mycoplasma pneumoniae evokes immunoglobulin M autoantibody which agglutinates human erythrocytes at 4°C (cold agglutinin) and is specific for I antigen. Cross-reactions between surface antigens of M. pneumoniae and human erythrocytes, previously examined by serological analysis, were examined by transmission and scanning electron microscopy. Ferritinlabeled human antimycoplasmal and rabbit antisera to erythrocyte membrane components reacted with antigens on the surface of both M. pneumoniae and erythrocytes. Adsorption of human erythrocytes to M. pneumoniae was blocked by the same antisera without ferritin label. It is proposed that the cross-reactive specificity lies in peripheral areas of the mycoplasmal cell, probably in a surface carbohydrate which has antigenic identity with erythrocyte glycoprotein.